Abstract: Methods, systems, and non-transitory computer-readable media are configured to perform operations comprising determining lane detection data and object detection data associated with an environment; determining a lane template based on the lane detection data; and generating localization data that identifies a location of an object in the environment based on the lane template and the object detection data.

Abstract: Methods, systems, and non-transitory computer-readable media are configured to perform operations comprising determining an image of an environment. A drivable area and boundary information associated with the environment are determined based on the image. At least one boundary for navigation of the environment is generated based on the drivable area and boundary information.

Abstract: A technique for threat response associated with an endpoint detection and response (EDR) system. The system uses a combination of automated observable detection, threat intelligence enrichment, graph analysis, and supervised machine learning to machine-predict analyst behavior in classifying (as ‘true’ or ‘false’ positives) the EDR alerts, and to support either (i) automated suppression of those alerts that the system classifies with sufficient confidence as either true or false, or (ii) for those alerts than cannot be so classified, the providing of recommendations to analysts to facilitate their activities. Auto-detection of observables for graph-based feature detection, together with the automated disposition of alerts where possible greatly reduces overall analyst workload for the EDR system.

Abstract: Systems, devices, computer-implemented methods, and tangible non-transitory computer-readable media for generating reports from one or more databases that store disparate datasets are provided. Specifically, the proposed systems enable the intelligent generation of reports from multiple datasets by automatically determining a proposed set of join configurations for combination of the multiple datasets. Unique identifiers can be assigned to a portion of the combined dataset and the combined dataset can be expanded for performing an aggregation operation associated with the portion of the combined dataset by referencing the one or more unique identifiers. Reports can be generated based on the combined dataset and the aggregation operation.

Abstract: This application is directed to augmenting training images used for generating vehicle driving models. A computer system obtains a first image of a road, identifies within the first image a drivable area of the road, obtains an image of a traffic safety object, and determines a detour path on the drivable area. The computer system determines positions of a plurality of traffic safety objects to be placed adjacent to the detour path, and generates a second image from the first image by adaptively overlaying a respective copy of the image of the traffic safety object at each of the positions of the plurality of traffic safety objects on the drivable area within the first image. The second image is added to a corpus of training images to be used by a machine learning system to generate a model for facilitating driving a vehicle (e.g., at least partial autonomously).

Abstract: Embodiments provide improved techniques for identifying object and vehicle locations on an HD map using camera images. Synthetic data may be generated by populating a high-definition (HD) map with an object. The populated HD map can be projected to a two-dimensional camera view image depicting the object. The HD map and object detection data generated from image (e.g., identifying lane/object locations within the image) can be used to train the model to identify HD map locations of an autonomous vehicle capturing images as it travels, as well as various objects detected from those captured images. Subsequently, a new image may be processed using object detection techniques to detect lane/object locations within the image. An HD map and the detected lane/object data may be provided to the model, which in turn identifies, on the HD map, the locations of the vehicle and the various objects.

Abstract: Certain aspects and features of this disclosure relate to semantically-aware image extrapolation. In one example, an input image is segmented to produce an input segmentation map of object instances in the input image. An object generation network is used to generate an extrapolated semantic label map for an extrapolated image. The extrapolated semantic label map includes instances in the original image and instances that will appear in an outpainted region of the extrapolated image. A panoptic label map is derived from coordinates of output instances in the extrapolated image and used to identify partial instances and boundaries. Instance-aware context normalization is used to apply one or more characteristics from the input image to the outpainted region to maintain semantic continuity. The extrapolated image includes the original image and the outpainted region and can be rendered or stored for future use.

Abstract: A system can access, from an interaction database, customer interaction information comprising a customer interaction, wherein the customer interaction information represents an interaction between a customer and an entity, and determine, based on a complaint identification model, whether a segment of the interaction comprises a complaint, wherein the complaint identification model has been generated based on machine learning applied to past customer interaction information representative of past customer interactions other than the customer interaction.

Abstract: This application is directed to augmenting training images used for generating vehicle driving models. A computer system obtains a first image of a road, identifies within the first image a drivable area of the road, obtains an image of a traffic safety object, and determines a detour path on the drivable area. The computer system determines positions of a plurality of traffic safety objects to be placed adjacent to the detour path, and generates a second image from the first image by adaptively overlaying a respective copy of the image of the traffic safety object at each of the positions of the plurality of traffic safety objects on the drivable area within the first image. The second image is added to a corpus of training images to be used by a machine learning system to generate a model for facilitating driving a vehicle (e.g., at least partial autonomously).

Abstract: A method of automatically producing maps and measures that visualize and quantify placement and progress of construction elements, such as walls, ducts, etc. in images. From a set of images depicting a scene, element confidences per pixel in each of the images are produced using a classification model that assigns such confidences. Thereafter, element confidences for each respective one of a set of 3D points represented in the scene are determined by aggregating the per-pixel element confidences from corresponding pixels of each of the images that is known to observe the respective 3D points. These element confidences are then updated based on primitive templates representing element geometry to produce a 3D progress model of the scene.

Abstract: A 360 panoramic video sequence of an architectural or industrial site includes a plurality of 360 images. 3D poses of the 360 images are determined with respect to one another, and a subset of the images extracted for further processing according to selection criteria. The 3D poses of the extracted images is refined based on determined correspondences of the features in the images, and a 3D point cloud of the site is developed from the extracted images with refined pose estimates. An “as-built” representation of the site captured in the panoramic video sequence is created from the sparse 3D point cloud and aligned to a 2D or 3D plan of the site. Once so aligned, the extracted images may be presented to a user in conjunction with their plan positions (e.g., overlayed on the site plan). Optionally, a point cloud or mesh view of the site may also be returned.

Abstract: A 360 panoramic video sequence of an architectural or industrial site includes a plurality of 360 images. 3D poses of the 360 images are determined with respect to one another, and a subset of the images extracted for further processing according to selection criteria. The 3D poses of the extracted images is refined based on determined correspondences of the features in the images, and a 3D point cloud of the site is developed from the extracted images with refined pose estimates. An “as-built” representation of the site captured in the panoramic video sequence is created from the sparse 3D point cloud and aligned to a 2D or 3D plan of the site. Once so aligned, the extracted images may be presented to a user in conjunction with their plan positions (e.g., overlayed on the site plan). Optionally, a point cloud or mesh view of the site may also be returned.

Abstract: A 360 panoramic video sequence of an architectural or industrial site includes a plurality of 360 images. 3D poses of the 360 images are determined with respect to one another, and a subset of the images extracted for further processing according to selection criteria. The 3D poses of the extracted images is refined based on determined correspondences of the features in the images, and a 3D point cloud of the site is developed from the extracted images with refined pose estimates. An “as-built” representation of the site captured in the panoramic video sequence is created from the sparse 3D point cloud and aligned to a 2D or 3D plan of the site. Once so aligned, the extracted images may be presented to a user in conjunction with their plan positions (e.g., overlayed on the site plan). Optionally, a point cloud or mesh view of the site may also be returned.

Abstract: A 360 panoramic video sequence of an architectural or industrial site includes a plurality of 360 images. 3D poses of the 360 images are determined with respect to one another, and a subset of the images extracted for further processing according to selection criteria. The 3D poses of the extracted images is refined based on determined correspondences of the features in the images, and a 3D point cloud of the site is developed from the extracted images with refined pose estimates. An “as-built” representation of the site captured in the panoramic video sequence is created from the sparse 3D point cloud and aligned to a 2D or 3D plan of the site. Once so aligned, the extracted images may be presented to a user in conjunction with their plan positions (e.g., overlayed on the site plan). Optionally, a point cloud or mesh view of the site may also be returned.